Table 3.

The performance of combined models constructed by six algorithms in the internal test set.

	AUC	95%CI	SEN	95%CI	SPE	95%CI	ACC	95%CI
KNN+ Clinical	0.605	0.507-0.611	0.777	0.677-0.853	0.341	0.245-0.451	0.566	0.491-0.639
LR+ Clinical	0.706	0.610-0.725	0.723	0.620-0.808	0.614	0.503-0.714	0.670	0.570-0.738
DT+ Clinical	0.662	0.538-0.655	0.511	0.406-0.614	0.670	0.561-0.765	0.588	0.513-0.660
Lin-SVM+ Clinical	0.709	0.634-0.786	0.702	0.600-0.790	0.636	0.526-0.734	0.670	0.600-0.738
Gaus-SVM+ Clinical	0.644	0.520-0.630	0.766	0.665-0.845	0.386	0.286-0.497	0.582	0.507-0.645
Ploy-SVM+ Clinical	0.663	0.533-0.647	0.691	0.587-0.780	0.489	0.381-0.597	0.593	0.518-0.666
Clinical	0.660	0.570-0.786	0.574	0.468-0.674	0.636	0.526-0.734	0.604	0.530-0.676

KNN, K-nearest neighbor; LR, logistic regression; DT, decision tree; Lin-SVM, linear-support vector machine; Gaus-SVM, Gaussian- support vector machine; Ploy-SVM, polynomial-SVM.